- · 《遥感学报》栏目设置[09/01]
- · 《遥感学报》数据库收录[09/01]
- · 《遥感学报》投稿方式[09/01]
- · 《遥感学报》征稿要求[09/01]
- · 《遥感学报》刊物宗旨[09/01]
来稿应自觉遵守国家有关著作权法律法规，不得侵犯他人版权或其他权利，如果出现问题作者文责自负，而且本刊将依法追究侵权行为给本刊造成的损失责任。本刊对录用稿有修改、删节权。经本刊通知进行修改的稿件或被采用的稿件，作者必须保证本刊的独立发表权。 一、投稿方式： 1、 请从 我刊官网 直接投稿 。 2、 请 从我编辑部编辑的推广链接进入我刊投审稿系统进行投稿。 二、稿件著作权： 1、 投稿人保证其向我刊所投之作品是其本人或与他人合作创作之成果，或对所投作品拥有合法的著作权，无第三人对其作品提出可成立之权利主张。 2、 投稿人保证向我刊所投之稿件，尚未在任何媒体上发表。 3、 投稿人保证其作品不含有违反宪法、法律及损害社会公共利益之内容。 4、 投稿人向我刊所投之作品不得同时向第三方投送，即不允许一稿多投。 5、 投稿人授予我刊享有作品专有使用权的方式包括但不限于：通过网络向公众传播、复制、摘编、表演、播放、展览、发行、摄制电影、电视、录像制品、录制录音制品、制作数字化制品、改编、翻译、注释、编辑，以及出版、许可其他媒体、网站及单位转载、摘编、播放、录制、翻译、注释、编辑、改编、摄制。 6、 第5条所述之网络是指通过我刊官网。 7、 投稿人委托我刊声明，未经我方许可，任何网站、媒体、组织不得转载、摘编其作品。
复杂卫星图像中的小目标船舶识别(5)
作者:网站采编关键词:
摘要：Lin Y O,Lei H,Li X Y and Wu learning in NLP:meth‐ods and of University of Electronic Science and Technology of China,46(6):913-919(林奕欧,雷航,李晓瑜,吴佳.2017.自然语言处理中的深度
Lin Y O,Lei H,Li X Y and Wu learning in NLP:meth‐ods and of University of Electronic Science and Technology of China,46(6):913-919(林奕欧,雷航,李晓瑜,吴佳.2017.自然语言处理中的深度学习:方法及应用.电子科技大学学报,46(6):913-919)[DOI:10.3969/]
Liu W,Anguelov D,Erhan D,Szegedy C,Reed S,Fu C Y and Berg A :single shot multibox detector//Proceedings of 14th European Conference on Computer ,The Nether‐lands:Springer:21-37[DOI:10.1007/978-3-319--0_2]
Liu Y,Hu D and Fan J L.2018.A survey of crime scene investigation image Electronica Sinica,46(3):761-768(刘颖,胡丹,范九伦.2018.现勘图像检索综述.电子学报,46(3):761-768)[DOI:10.3969/]
Rafique M A,Pedrycz W and Jeon license plate de‐tection using region-based convolutional neural Computing,22(19): 6429-6440 [DOI:10.1007/s00500-017-2696-2]
Redmon J,Divvala S,Girshick R and Farhadi only look once:unified,real-time object detection//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Vegas,NV,USA:IEEE:779-788[DOI:10.1109/]
Ren S Q,He K M,Girshick R and Sun R-CNN:towards real-time object detection with region proposal Transactions on Pattern Analysis and Machine Intelligence,39(6):1137-1149[DOI:10.1109/]
Szegedy C,Liu W,Jia Y Q,Sermanet P,Reed S,Anguelov D,Erhan D,Vanhoucke V and Rabinovich deeper with convo‐lutions//Proceedings of 2015 IEEE Conference on Computer Vi‐sion and Pattern ,MA,USA:IEEE:1-9[DOI:10.1109/]
Tang Z T,Shao K,Zhao D B and Zhu Y progress of deep reinforcement learning:from AlphaGo to AlphaGo Theory and Applications,34(12):1529-1546(唐振韬,邵坤,赵冬斌,朱圆恒.2017.深度强化学习进展:从AlphaGo到AlphaGo Zero.控制理论与应用,34(12):1529-1546)[DOI:10.7641/]
Tian J X,Liu G C,Gu S S,Ju Z J,Liu J G and Gu D learning in medical image analysis and its Auto‐matica Sinica,44(3):401-424(田娟秀,刘国才,谷珊珊,鞠忠建,刘劲光,顾冬冬.2018.医学图像分析深度学习方法研究与挑战.自动化学报,44(3):401-424)[DOI:10./]
Vedaldi A and Lenc :convolutional neural net‐works for MATLAB//Proceedings of the 23rd ACM international conference on ,Australia:ACM:689-692[DOI:10.1145/.]
Wang P Y,Li F,Zhou S R and Liao Z F.2017.A ship detection and tracking algorithm under complex wharf Engineering and Science,39(5):992-998(王培玉,李峰,周书仁,廖卓凡.2017.复杂码头环境下的船舶检测与跟踪算法.计算机工程与科学,39(5):992-998)[DOI:10.3969/]
Wei application of template matching and BP neural net‐work in ship Science and Technology,38(20):133-135(魏娜.2016.模板匹配和BP神经网络在船舶识别中的应用.舰船科学技术,38(20):133-135)[DOI:10.3404/]
Xiao Z on ship identification and control terminal system based on active RFID Science and Tech‐nology,39(8A):142-144(肖志良.2017.基于有源RFID技术的船舶识别与控制终端系统研究.舰船科学技术,39(8A):142-144)[DOI:10.3404/]
Yang M,Ruan Y D,Chen L K,Zhang P and Chen Q vid‐eo recognition algorithms for inland river ships based on faster of Beijing University of Posts and Telecommunica‐tions,40(S1):130-134(杨名,阮雅端,陈林凯,张鹏,陈启美.2017.甚高速区域卷积神经网络的船舶视频检测方法.北京邮电大学学报,40(S1):130-134)[DOI:10./]
Zhang M M,Choi J,Daniilidis K,Wolf M T and Kanan :a dataset for recognizing maritime imagery in the visible and infra‐red spectrums//Proceedings of 2015 IEEE Conference on Comput‐er Vision and Pattern Recognition ,MA,USA:IEEE:10-16[DOI:10.1109/]
Zhou Y and Zhang J deep learning for product im‐age of Computer Research and Development,54(8):1824-1832(周晔,张军平.2017.基于多尺度深度学习的商品图像检索.计算机研究与发展,54(8):1824-1832)[DOI:10.7544/]
First,this work uses negative sample enhancement learning to train the model,and fog and coastal backgrounds are sent as negative samples to the network for training to solve the influence of complex sea conditions,such as cloud-fog occlusion,and coastal the same time,a multiscale sample training method is used in this paper in view of the problem that the size of some targets in the image is small and affects recognition images are compressed into multiple scales and sent to the network for training so that the network can fully learn the features of various ship sizes,thereby solving the difficulty of small target ,the pre-trained ZF model is used for feature extraction,and the feature maps are sent to the region proposal network to generate proposal ,the generated candidate areas are sent to the fully connected layer for ship fine-grained recognition.
Experimental results show that the precision and recall of our method increased by 6.98%and 18.17%respectively,and the accuracy of ship recognition can reach 92.27%compared with Faster method can guarantee real-time requirement based on high recognition accuracy and recognize ships under various conditions.
The trained model can realize the fine-grained recognition of model not only solves the problems of cloud-fog occlusion to ships,but also the difficulty of small target accuracy and real-time performance of the model meet the actual requirements and has strong ,the experimental results also show that the method used still has ,the constructed network structure has high complexity and excessive network overhead,which increase the processing time of ship target recognition;Secondly,the accuracy of the trained ship recognition model can still be two points are also the key tasks for the follow-up work.
文章来源：《遥感学报》 网址: http://www.ygxbzz.cn/qikandaodu/2021/0409/597.html